Parochialism, social norms, and discriminationagainst immigrantsDonghyun Danny Choia,1,2, Mathias Poertnerb,1,2, and Nicholas Sambanisa,1,2

aDepartment of Political Science, University of Pennsylvania, Philadelphia, PA 19106; and bEvidence in Governance and Politics, University of California,Berkeley, CA 94720

Edited by Ryan D. Enos, Harvard University, Cambridge, MA, and accepted by Editorial Board Member Mary C. Waters May 2, 2019 (received for reviewNovember 26, 2018)

Ingroup bias and outgroup prejudice are pervasive features ofhuman behavior, motivating various forms of discrimination andconflict. In an era of increased cross-border migration, thesetendencies exacerbate intergroup conflict between native pop-ulations and immigrant groups, raising the question of howconflict can be overcome. We address this question through alarge-scale field intervention conducted in 28 cities across threeGerman states, designed to measure assistance provided to immi-grants during everyday social interactions. This randomized trialfound that cultural integration signaled through shared socialnorms mitigates—but does not eliminate—bias against immi-grants driven by perceptions of religious differences. Our resultssuggest that eliminating or suppressing ascriptive (e.g., ethnic)differences is not a necessary path to conflict reduction in mul-ticultural societies; rather, achieving a shared understanding ofcivic behavior can form the basis of cooperation.

immigration | discrimination | cultural integration | norms |field experiment

Parochialism—the tendency to favor ingroup members at theexpense of an outgroup—has been identified by observa-

tional and experimental research in the social and evolutionarysciences as one of the fundamental tenets of human behavior(1–6). Ingroup bias and outgroup prejudice—two componentsof parochialism—are pervasive features of intergroup relationsacross cultures and have been tied to various forms of con-flict, ranging from discrimination in the labor market (7, 8),to racial profiling in criminal justice (9), to suicide bombings(10) and mass atrocities such as genocide (11). Parochialismhas been shown to coevolve with conditions that favor inter-group conflict (2, 12, 13). Cultural group formation can beexplained as a mechanism for a population of heterogeneousindividuals to resolve coordination problems (14) so that theycan compete more effectively for resources and survival. Amongearly humans, war solidified altruism toward members of one’singroup (12, 15). In contemporary history, conflicts along ethno-religious lines have created salient national identities and havepolarized populations.

Any conflict over economic resources, power, or identity couldheighten the salience of ingroup–outgroup boundaries, increas-ing parochialism. Large-scale cross-border immigration can gen-erate such conflicts. Although immigration can be beneficial byaddressing labor market needs and forging new networks acrosssocieties, it also challenges native social norms and forces soci-eties to adapt to rapid population change. Wars in Syria, Iraq,Afghanistan, and elsewhere have caused the largest refugee crisissince World War II, leading to a sharp rise in immigration frommajority Muslim countries to Europe. Populist politicians andmany of the media discourses across Europe have emphasizedethnic, religious, and cultural differences between immigrantsand the majority population in host countries to mobilize publicopinion against immigration.

A survey-based literature in political science has documentedopposition to immigration (16, 17) deriving from economiccompetition between native and immigrant groups (18–20) as

well as a perceived cultural threat (21) and a fear of a mul-ticultural future (22, 23). The idea that discrimination againstimmigrant populations is by and large a “cultural” phenomenonis no longer contested (ref. 16, p. 231). Whereas competingeconomic interests between native populations and immigrantscan account for anti-immigrant hostility, cultural differences areoften at the core of anti-immigrant attitudes (e.g., refs. 24–26).Grounded in seminal theories of social identity (27), prejudice(28, 29), and ethnocentrism (30), many of these studies regardsentiments toward immigrants as a manifestation of the hostpopulation’s ingroup identity and the extent to which immigrantgroups are perceived to be “distinct,” and therefore “distant,”from their own (31–34). Such “otherness” of immigrant groupscan induce natives to develop prejudices and stereotypes thatultimately culminate in negative attitudes and predispositionstoward immigrants, whom they consider to pose a sociotropicthreat to their own group (ref. 16, p. 232). Both recent immi-grant groups and other minority populations of immigrant back-ground are vulnerable to the consequences of this “othering”process.

Despite the consensus on the centrality of cultural expla-nations, there is a dearth of investigations that examine whattypes of cultural distinctions between native and immigrantgroups most critically influence native attitudes toward immigra-tion (ref. 16, p. 242). Immigrant groups often possess a diverse

Significance

As immigration increases due to globalization, wars, and cli-mate change, there is more interaction between native andimmigrant populations and more potential for conflict. Pol-icymakers have emphasized the need to better integrateimmigrants to forge a common set of norms concerningappropriate behavior. However, there is a lack of evidence tosupport the hypothesis that a shared understanding of normscan reduce discrimination. We provide experimental evidencethat religious differences cause bias and discrimination ineveryday interactions between natives and immigrants. Cul-tural integration signaled through immigrants’ enforcementof local norms reduces, but does not eliminate, bias. As longas public debates and policies heighten the importance ofreligious differences, cultural integration will not be able toeliminate intergroup conflict.

D.D.C, M.P., and N.S. designed research, performed research, analyzed data, and wrotethe paper.y

The authors declare no conflict of interest.y

This article is a PNAS Direct Submission. R.D.E. is a guest editor invited by the EditorialBoard.y

Published under the PNAS license.y1 D.D.C., M.P., and N.S. contributed equally to this work.y2 To whom correspondence may be addressed. Email: dhchoi@sas.upenn.edu, mathias.poertner@berkeley.edu, or sambanis@upenn.edu.y

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1820146116/-/DCSupplemental.y

Published online July 29, 2019.

16274–16279 | PNAS | August 13, 2019 | vol. 116 | no. 33 www.pnas.org/cgi/doi/10.1073/pnas.1820146116

Dow

nloa

ded

by g

uest

on

Oct

ober

9, 2

020

POLI

TICA

LSC

IEN

CES

array of attributes that can separately distinguish them fromhost populations. Questions remain as to whether each of theseimmigrant attributes contributes equally (or at least similarly)toward determining the perceived distance between native andimmigrant populations.

Only recently have scholars begun to isolate the impact ofdifferent immigrant attributes on native population attitudes.Adida, Laitin, and Valfort (8), for example, use an audit exper-iment to examine whether Muslim immigrants are likely to facediscrimination in the French labor market. The authors findthat holding immigrant ethnicity constant, job applicants withputatively Muslim names are much less likely to receive a jobinterview callback. These results confirm lessons learned froma large literature on ethnic discrimination in the market (35,36). Hainmueller and Hopkins (33) parse out the relative weightof different immigrant attributes in the United States to findthat an immigrant’s labor market qualifications such as level ofeducation and language skills, as well as an immigrant’s coun-try of origin, affect hypothetical decisions about immigrationadmission.

Big questions remain unaddressed or underexplored. Forexample, while studies have found that immigrants from cer-tain countries are preferred by natives (33, 37), it is less clearwhat about the information conveyed in the immigrant’s countryof origin is influencing immigration attitudes. Does the countryof origin evoke prejudices against individuals that appear visu-ally different from native populations? Or does it make salientthe perceived biases against groups that host populations con-sider to be socio-culturally incompatible with their own? We lacka full understanding of whether anti-immigrant sentiments areprimarily driven by ethnic attributes such as skin tone, phenotyp-ical, or religious difference or other qualities that are consideredto be under the control of the immigrants themselves, such asthe immigrants’ behavior and the degree of cultural integration.If indeed ethnic attributes are important, it remains to be con-vincingly established whether such effects can be counteractedby the immigrants’ acquisition of attributes and skills that signalwillingness to integrate into the native society.

Our study targets these gaps in the literature, using a fieldexperiment. While significant efforts have been directed towardcapturing immigration attitudes outside the context of surveysor laboratory environments, most of these studies have focusedon the labor market or other economic domains (2, 37). Whatwe need is a deeper understanding of why different immigrantgroups are treated differently, a question that we are able toaddress by focusing on how norm assimilation and religiousdifference affect behavior toward the same immigrant. Thisquestion is distinct from the question of whether native popu-lations discriminate against immigrants. Previous experimentalstudies have presented evidence of discrimination against immi-grants (38, 39). Winter and Zhang (40) have shown that Germannatives are more likely than immigrant minorities to enforcenorms and that minorities are more likely to be sanctionedfor norm violations than are natives. Our study advances ourunderstanding of the determinants of behavior toward differ-ent types of immigrants and helps us consider how to reducediscrimination in typical day-to-day social interactions betweennative and immigrant populations in the real world. Insofar asnative population preferences toward immigration and immi-gration policy reflect the realization of latent attitudes thatnatives form in the process of engaging in day-to-day interactionswith immigrants, failure to analyze these “everyday” behaviorsis an important oversight. Our study is designed to addressthis void with a field experiment that measures discriminationamong individuals who are unaware that their behavior is beingobserved.

Our research design disentangles the effects of differentascriptive characteristics that can generate ingroup bias and goes

farther than previous studies by exploring whether parochialismcan be overcome by cultural integration. Socio-biological andpsychological studies of parochialism have established that ahuman neurobiological architecture evolved to highlight differ-ences between “us” and “them.” That architecture is not compat-ible with the growing multiculturalism of our societies (41) unlessgovernments or powerful societal actors take steps to rede-fine the ingroup by integrating outgroups such as immigrants.To reduce intergroup conflict, policymakers have emphasizedthe need to better integrate immigrants not only economically,but also culturally in native societies, to forge a common setof rules and norms concerning the boundaries of appropriatebehavior. However, there is a lack of empirical evidence tosupport the hypothesis that a shared understanding of socialnorms can reduce prejudice and discrimination. In particular,real-world experimental interventions addressing these issues arerare (42).

We provide experimental evidence on whether cultural inte-gration through the enforcement of common social norms canreduce bias against immigrants driven by perceptions of ascrip-tive (religious) differences. By “cultural integration,” we meanthe adoption of mutually accepted standards of behavior. Suchagreement could be the result of acculturation or assimila-tion (adopting the norms of the native society) or the prod-uct of dialogue and fusion between the native society andimmigrant communities over time. To test whether culturalintegration signaled through norm enforcement reduces discrim-ination against immigrants, we implemented a field interven-tion in Germany, which is currently experiencing one of thelargest waves of immigration in modern European history; since2015, more than 1.5 million individuals have applied for asy-lum in Germany, making it the largest recipient of refugeesin Europe (43). Immigration has emerged as a salient issuein German politics, making Germany an ideal setting in whichto conduct our study. Germans of immigrant background arealso affected by the backlash to the refugee crisis as the cri-sis has sparked debates about the future of multiculturalismin Europe.

Research DesignWe created a highly realistic and carefully controlled “microen-vironment” (44) of intergroup social interactions designed toobserve the degree of assistance offered to strangers, some ofwhom were German and others immigrants. The experiment,depicted in Fig. 1, involved the staged violation of the normagainst littering in public spaces* followed by an interactionthat was designed to elicit behavioral responses by unknowingbystanders (see SI Appendix, Figs. S1 and S2 for a schematicrepresentation of the experiment). The intervention proceededas follows: A male German confederate (violator) violates awidely held norm against littering by dropping an empty cof-fee cup on a train platform. A female confederate (punisher)immediately sanctions the violator by asking him to pick uphis trash. The violator unwillingly but promptly complies withthe request and leaves the scene. As the violator is walk-ing away, the punisher receives a call in view of bystanderswho have witnessed this previous interaction. While engagedin the call, she accidentally drops a bag of her possessions,the contents (oranges) of which disperse on the platform.The confederate appears to be in need of help retrieving herpossessions, and we measure whether bystanders provide thatassistance.

*Balafoutas et al. (45, 46) use a related approach to measure direct vs. indirect punish-ment of norm violators and whether altruistic punishment covaries with the severerityof norm violations. We use parts of their design and extend it to address differentquestions.

Choi et al. PNAS | August 13, 2019 | vol. 116 | no. 33 | 16275

Dow

nloa

ded

by g

uest

on

Oct

ober

9, 2

020

Fig. 1. Experiment in progress. (A) In the first step, the male confederateviolates the norm against littering by discarding an empty coffee cup. (B)The female confederate promptly sanctions the male confederate for thenorm violation. (C) The female confederate then drops personal possessionsand is in need of assistance.

Two key dimensions of the intervention were manipulatedexperimentally. First, to vary the female confederate’s perceivedmembership in the ingroup (natives) or outgroup (immigrants),we randomly assigned the ethnoreligious attributes of the con-federate to one of four conditions: an immigrant wearing a hijab,the same immigrant in plain attire without a hijab, the sameimmigrant wearing a clearly visible Christian cross, or a Ger-man female, who would serve as our control condition (skin toneand phenotype were the variables that would help the bystandersidentify confederates as immigrants in the control condition).Second, we also manipulated the level of cultural integration byrandomly varying whether the female confederate enforced theantilittering norm. This action signaled to bystanders that theconfederate shared their norms and was a civic-minded person.In roughly half of our sample, the female confederate sanctionedthe norm violator before requiring assistance. In the remain-ing half, a different confederate enforced the norm instead(see Fig. 2 for the matrix of treatment conditions). Apart fromthese two dimensions, we controlled for social class by havingconfederates wear similar attire across the different teams anditerations. We minimized the potential for differences in attrac-tiveness to affect assistance rates by having the same minorityconfederate (immigrant female) play all three roles. Since it isnot possible for the same actor to portray immigrant and native,we rotated the person playing the role of the German femaleconfederate in each team. In SI Appendix, we also show that ourresults hold using team fixed effects, which analyze within-teamvariation in assistance rates across iterations.

One design feature of the intervention warrants note: Thedecision to manipulate the religious characteristics of the con-federate in addition to her ethnic attributes stemmed from thecentral role that religious difference plays in political debateson immigration in Germany and Europe (Fig. 3). Given mediacoverage of cultural conflicts between European Christian and

immigrant Muslim populations, we had reason to believe thatreligious identity would be central for the delineation of ingroupsand outgroups in Germany.

The interventions were conducted in 29 train stations acrossthree German states (North Rhine-Westphalia, Saxony, andBrandenburg). We implemented a total of 1,614 iterations of theintervention, involving 7,142 bystanders, in major train stationsof these cities over a 3-wk period between July and August 2018.During each iteration, confederates were tasked with record-ing the behavior of bystanders who observed the intervention(coders were not blinded; see SI Appendix for more discussion).The main outcome of interest, which was coded at the itera-tion level, was whether any bystanders offered assistance to thefemale confederate in retrieving her possessions. Confederatesalso noted the total number and gender of bystanders withina prespecified radius, as well as other characteristics of eachiteration.

The research protocol was reviewed and approved by Univer-sity of Pennsylvania’s Institutional Review Board (IRB Protocol829824). A waiver of the consent process was obtained (seeSI Appendix for additional information on ethical and safetyconsiderations).

ResultsSI Appendix provides further details regarding the design of theintervention, timelines for the study, data collection locations,and protocols for data collection, as well as auxiliary analyses.Auxiliary analyses in SI Appendix include regression-based anal-ysis of treatment effects including state fixed effects, bystanderfixed effects, and various controls; comparison of effects inEast vs. West Germany; effects of language assimilation on biastoward immigrants; and an online survey to explore Germans’attitudes toward littering and provide manipulation checks forsome of our treatments. All analyses presented in the maintext, unless otherwise noted, were prespecified in an analysisplan registered with the Evidence in Governance and Politics(EGAP) network (20180725AB) before commencement of datacollection.

We expected that immigrants would receive less assistancethan natives and that identity markers (hijabs) that increasethe difference (social distance) between natives and immigrantswould decrease assistance. Furthermore, since enforcement oflocal social norms regarding appropriate behavior signals cul-tural integration, we hypothesized that enforcing the antilitteringnorm would offset the negative bias toward immigrants. Wechose this particular norm because it constitutes a noncontrover-sial civic norm that is particularly widespread and deeply inter-nalized in Germany (see SI Appendix, section 7, for suggestiveevidence consistent with this assumption).

Fig. 2. Treatment assignment matrix. This matrix presents the eight experi-mental treatment and control conditions corresponding to two confederateethnoracial categories, three immigrant religious attribute categories, andtwo norm enforcement conditions.

16276 | www.pnas.org/cgi/doi/10.1073/pnas.1820146116 Choi et al.

Dow

nloa

ded

by g

uest

on

Oct

ober

9, 2

020

POLI

TICA

LSC

IEN

CES

Fig. 3. Manipulation of confederate identity.

First, our analyses provide strong evidence in support of ourhypothesis of bias against immigrants and suggest that religiousdifference is what defines immigrants as an outgroup. FemaleGerman confederates, who serve as our controls, were assistedin retrieving their possessions in 78.3% of all iterations. By con-trast, confederates of an immigrant background (immigrant withhijab and immigrants in plain attire) were assisted less, at 71.3%of iterations (see SI Appendix, Table S11 for further discussionof these analyses using regression with state-fixed effects andbystander fixed effects). The difference between the level ofassistance offered to immigrants vs. that to natives is thereforearound 7 percentage points and is statistically distinguishable atconventional levels (t = −2.11, P < 0.05, two-tailed). The resultsshown in Fig. 4 pool across other conditions (i.e., among normenforcers and nonenforcers within the native and immigrantgroups).

The negative bias against immigrants, however, is only due tointergroup differences in religious identity. Phenotypical (eth-nic) differences on their own are not sufficient to cause bias.As shown in column 3 of Fig. 4, immigrant confederates wear-ing a hijab, which clearly signals that they are of Muslim faith,were assisted only 66.3% of the time (these results pool acrossother treatment conditions, including norm enforcement). Thisis 12 percentage points less than assistance offered to Germanconfederates (t = −3.22, P < 0.001, two-tailed). In the sub-sample of interventions from the former East German state ofSaxony, which was the site of violent far-right anti-immigrantprotests in August 2018, this differential increases to almost22 percentage points (SI Appendix, Table S9). Differences indiscrimination in East vs. West Germany are explored in SIAppendix, Table S7, where we show that bias is larger in theEast, although we cannot establish the cause of these differ-ences. We also find that the rise in assistance levels due togood citizenship (i.e., prosocial behavior toward immigrants dueto norm enforcement) is significantly larger in the East (SIAppendix, Table S15). Economic differences between East andWest, the legacy of communism, as well as differences in religios-ity and in the degree of contact with immigrants could explainthose results. The magnitude of this negative bias is especiallynoteworthy given the nature of the items dropped in the inter-vention; the oranges dispersed in a manner that made it seemchallenging for our confederates to retrieve them expedientlyby themselves, which should have created strong pressures forbystanders to offer assistance regardless of the identity of theconfederate.

Our analysis provides corroborating evidence for the central-ity of religious differences as the basis of parochialism in ourcontext. Specifically, the average rate of assistance to immigrantconfederates in the control condition (column 3 in Fig. 4) isstatistically indistinguishable from the assistance given to nativeGermans. Thus, we do not find any evidence of ethnically drivenracism or discrimination in the context of our experiment. This isdespite the fact that bystanders recognize our immigrant confed-erates as non-German (see SI Appendix, section 6, for evidence

from a follow-up survey). Racial and phenotypical differencesalone are not sufficient to induce discrimination in the con-text of minor everyday interactions in our sample. That mightbe due to the fact that immigrants in the control conditionare dressed similarly to the native confederates and thereforesignal some degree of cultural integration. Immigrants who weara Christian cross (column 2 in Fig. 4) also send a signal of cul-tural integration. The difference between the level of assistanceoffered to a German confederate and that to the cross-wearingimmigrant confederate is a mere 1.96 percentage points and isstatistically indistinguishable from zero (t = 0.583, P = 0.56,two-tailed). However, there is a decrease of around 10 percent-age points in the assistance offered to the immigrant wearinga hijab relative to the immigrant control group or the immi-grant group wearing the cross (t = 3.29, P < 0.01, two-tailed).Given that the level of assistance offered to immigrants in thecontrol and in the cross condition is roughly the same as the assis-tance offered to natives, we can conclude that negative attitudestoward Muslim immigrants are what drive the results in Fig. 4and we find no evidence that racial or ethnic differences alonegenerate discrimination. In SI Appendix, Table S5, we conductanalyses using alternate outcome measures and show consis-tent results with respect to the share of bystanders who provideassistance.

Can a shared understanding of good citizenship—demon-strated through an immigrant’s enforcement of local socialnorms—help partially counteract the bias against Muslimimmigrants? Analyses presented in Fig. 5 show that it can.In the first step, we compare between-treatment conditionswherein an immigrant confederate wearing a hijab enforces theantilittering norm (column 3) and conditions in which she doesnot (column 4); the mean level of assistance provided to immi-grant enforcers is more than 12 percentage points higher thanthat to immigrant nonenforcers (we compare only native Ger-mans to immigrants with a hijab in Fig. 5) and is statisticallysignificant at conventional levels (t = 2.772, P < 0.01, two-tailed).

The magnitude of the offsetting effects of norms is clearlydemonstrated when we examine the difference in assistancerates for immigrant enforcers and native nonenforcers. As

78.3 76.4 76.4

66.3

1.9%p (p=0.56)−0.03%p (p=0.991)

10.0%p (P<0.01)

12.0%p (P<0.001)

0

30

60

90

(1)Native

(2)Immigrantw/cross

(3)Immigrant

control

(4)Immigrant

w/hijab

Ass

ista

nce

Rat

es (%

)

Fig. 4. Parochialism in the level of assistance offered to strangers. Barsrepresent the mean rates of assistance for the treatment conditions. Theerror bars present 95% confidence intervals for the means. The bracketsand accompanying information report results of a standard two-tailed dif-ference in means test of treatment conditions with P values in parentheses.Bystanders provided significantly more assistance to perceived natives thanto immigrants.

Choi et al. PNAS | August 13, 2019 | vol. 116 | no. 33 | 16277

Dow

nloa

ded

by g

uest

on

Oct

ober

9, 2

020

83.9

73.3 72.9

60.4

10.6%p (p=0.053)

0.4%p (p=0.926)

12.4%p (p=0.006)

0

30

60

90

(1)Germanenforcer

(2)German

non−enforcer

(3)Immigrant w/hijab

enforcer

(4)Immigrant w/hijab

non−enforcer

Ass

ista

nce

Rat

es (%

)

Fig. 5. Offsetting effects of norm enforcement on bias. Bars represent themean rates of assistance for the treatment conditions. The error bars present95% confidence intervals for the means. The brackets and accompanyinginformation report results of a standard two-tailed difference in means testof treatment conditions with P values in parentheses.

columns 2 and 3 in Fig. 5 demonstrate, norm enforcement bringsthe assistance provided to an immigrant confederate within0.4 percentage points of that to a native who did not enforce thenorm; the difference is statistically indistinguishable from zero(t = 0.093, P = 0.926).

While the evidence presented above highlights the capacity ofcultural integration to counteract bias against immigrants gener-ated by ascriptive (religious) differences, it nonetheless suggeststhat norm enforcement is unable to eliminate the bias in itsentirety; the t test between assistance offered to native (column1 in Fig. 5) and immigrant (column 3 in Fig. 5) enforcers revealsa substantial difference of 11 percentage points, which is sta-tistically significant at the P < 0.05 level (t = 2.211, P < 0.05,two-tailed).

DiscussionA common social identity can serve as the foundation of demo-cratic citizenship. Nationalism often creates that common iden-tity, but in an increasingly multicultural world, adherence tonational identities is supplanted by parochial attachments. Ourstudy, set against the backdrop of increasing intergroup con-flict between native and immigrant populations in Germany,was designed to measure the effect of shared norms on dis-crimination against immigrants and minorities of immigrant

background. Theories of social psychology and political sciencesuggest that a common set of norms can unify individuals whowould otherwise be divided by their ethnic or religious identities.Importantly, we explore voluntary adherence to norms ratherthan forced assimilation by immigrants into the native society.Norms need not reflect the one-sided imposition of cultural prac-tices that immigrant communities find foreign or threatening totheir own identities; rather, they could emerge as the result of adialogue and mutual adaptation by host and immigrant commu-nities. Our experiment is focused on measuring responsivenessto the norm against littering—a norm that is both broadly sharedamong Germans and not inherently antithetical to cultural val-ues or religious practices of immigrant groups. Adherence to thisnorm by immigrants signals that they care about their local envi-ronment and that they consider themselves part of their Germancommunities.

Our experimental evidence suggests that eliminating or sup-pressing ascriptive differences is not a necessary path to conflictreduction in multicultural societies. We find no evidence of eth-nic discrimination per se. The majority of the subjects in ourexperiment do behave in a cooperative manner toward bothGermans and immigrants and shared norms of civic behaviorreduce the differences in the level of assistance offered to thetwo groups. Our findings suggest that norms can form the basisfor the reduction in discrimination and improved cooperation.We present evidence that good citizenship among socially distantimmigrant minorities is not discounted by the majority popula-tion. Yet we also find that the effectiveness of norms in forgingintegration is constrained by the salience of intergroup differ-ences. In the German context, religious difference increasessocial distance between native and immigrant populations, butthat distance is not insurmountable. Public policies and politicalrhetoric have heightened the salience of religious markers so thatcultural practices that are perceived foreign increase conflict.It is an open question whether societies will evolve to embracethe pluralism that is the inevitable consequence of ongoingmacrolevel processes of globalization. Leaders and opinion mak-ers could help by redirecting the public’s attention from theascriptive differences that divide groups to the importance ofcultural integration as the way to promote cooperation.

ACKNOWLEDGMENTS. We thank Loukas Balafoutas, Carsten De Dreu,Eugen Dimant, Thad Dunning, Don Green, Guy Grossman, Dan Hopkins,Brendan O’Leary, Eunji Kim, Dorothy Kronick, Marc Meredith, NikosNikiforakis, Anne Norton, Paul Sniderman, and Nan Zhang; participants atthe Penn Identity and Conflict Laboratory Working Group; Workshop forNorms and Behavioral Change (NoBec) 2018 at the University of Pennsylva-nia; and the Cultural Transmission and Social Norms (CTSN 3) workshop atMassachusetts Institute of Technology for valuable comments and sugges-tions. We are also grateful to our excellent team of 34 confederates andenumerators for their assistance in the implementation of the experiment.

1. C. K. W. De Dreu et al., The neuropeptide oxytocin regulates parochial altruism inintergroup conflict among humans. Science 328, 1408–1411 (2010).

2. S. Bowles, Did warfare among ancestral hunter-gatherers affect the evolution ofhuman social behaviors? Science 324, 1293–1298 (2009).

3. R. D. Enos, Causal effect of intergroup contact on exclusionary attitudes. Proc. Natl.Acad. Sci. U.S.A. 111, 3699–3704 (2014).

4. G. Bornstein, Intergroup conflict: Individual, group, and collective interests. Pers. Soc.Psychol. Rev. 7, 129–145 (2003).

5. H. Bernhard, U. Fischbacher, E. Fehr, Parochial altruism in humans. Nature 442, 912–915 (2006).

6. H. Tajfel, M. G. Billig, R. P. Bundy, C. Flament, Social categorization and intergroupbehaviour. Eur. J. Soc. Psychol. 1, 149–178 (1971).

7. M. Bertrand, S. Mullainathan, Are Emily and Greg more employable than Lakishaand Jamal? A field experiment on labor market discrimination. Am. Econ. Rev. 94,991–1013 (2004).

8. C. L. Adida, D. D. Laitin, M.-A. Valfort, Identifying barriers to Muslim integration inFrance. Proc. Natl. Acad. Sci. U.S.A. 107, 22384–22390 (2010).

9. K. Antonovics, B. G. Knight, A new look at racial profiling: Evidence from the Bostonpolice department. Rev. Econ. Stat. 91, 163–177 (2009).

10. S. Atran, J. Ginges, Religious and sacred imperatives in human conflict. Science, 336,855–857 (2012).

11. J. A. C. Everett, N. S. Faber, M. Crockett, Preferences and beliefs in ingroup favoritism.Front. Behav. Neurosci. 9, 15 (2015).

12. S. Bowles, J.-K. Choi, The coevolution of parochial altruism and war. Science 318,636–640 (2007).

13. S. Bowles, Warriors, levelers, and the role of conflict in human social evolution.Science 336, 876–879 (2012).

14. C. Efferson, R. Lalive, E. Fehr, The coevolution of cultural groups and ingroupfavoritism. Science 321, 1844–1849 (2008).

15. S. Bowles, Group competition, reproductive leveling, and the evolution of humanaltruism. Science 314, 1569–1572 (2006).

16. J. Hainmueller, D. J. Hopkins, Public attitudes toward immigration. Annu. Rev. Polit.Sci. 17, 225–249 (2014).

17. K. Bansak, J. Hainmueller, D. Hangartner, How economic, humanitarian, and religiousconcerns shape European attitudes toward asylum seekers. Science 354, 217–222(2016).

18. K. F. Scheve, M. J. Slaughter, Labor market competition and individual preferencesover immigration policy. Rev. Econ. Stat. 83, 133–145 (2001).

19. A. M. Mayda, Who is against immigration? A cross-country investigation of individualattitudes toward immigrants. Rev. Econ. Stat. 88, 510–530 (2006).

20. R. M. Dancygier, M. J. Donnelly, Sectoral economies, economic contexts, and attitudestoward immigration. J. Polit. 75, 17–35 (2013).

16278 | www.pnas.org/cgi/doi/10.1073/pnas.1820146116 Choi et al.

Dow

nloa

ded

by g

uest

on

Oct

ober

9, 2

020

POLI

TICA

LSC

IEN

CES

21. P. M. Sniderman, L. Hagendoorn, When Ways of Life Collide: Multiculturalism and itsDiscontents in the Netherlands (Princeton University Press, Princeton, NJ, 2007).

22. J. Sides, J. Citrin, European opinion about immigration: The role of identities,interests, and information. Br. J. Polit. Sci. 37, 477–504 (2007).

23. P. M. Sniderman, P. Pierangelo, R. de Figueiredo, T. Piazza, The Outsider: Prejudiceand Politics in Italy (Princeton University Press, Princeton, NJ, 2002).

24. P. M. Sniderman, L. Hagendoorn, M. Prior, Predisposing factors and situational trig-gers: Exclusionary reactions to immigrant minorities. Am. Polit. Sci. Rev. 98, 35–49(2004).

25. J. Hainmueller, M. J. Hiscox, Attitudes toward highly skilled and low-skilled immi-gration: Evidence from a survey experiment. Am. Polit. Sci. Rev. 104, 61–84(2010).

26. D. Tingley, Public finance and immigration preferences: A lost connection? Polity 45,4–33 (2012).

27. H. Tajfel, Human Groups and Social Categories: Studies in Social Psychology(Cambridge University Press, New York, NY, 1981).

28. G. W. Allport, The Nature of Prejudice (Basic Books, New York, NY, 1979).29. E. L. Paluck, D. P. Green, Prejudice reduction: What works? A review and assessment

of research and practice. Annu. Rev. Psychol. 60, 339–367 (2009).30. D. R. Kinder, C. D. Kam, Us against Them: Ethnocentric Foundations of American

Opinion (University of Chicago Press, Chicago, IL, 2010).31. C. Dustmann, I. P. Preston. Racial and economic factors in attitudes to immigration.

The BE Journal of Economic Analysis & Policy 7, 62 (2007).32. D. J. Schildkraut, Americanism in the Twenty-First Century: Public Opinion in the Age

of Immigration (Cambridge University Press, New York, NY, 2010).33. J. Hainmueller, D. J. Hopkins, The hidden American immigration consensus: A

conjoint analysis of attitudes toward immigrants. Am. J. Polit. Sci. 59, 529–548 (2015).

34. R. Maxwell, Occupations, national identity, and immigrant integration. Comp. Polit.Stud. 50, 232–263 (2017).

35. E. Zschirnt, D. Ruedin, Ethnic discrimination in hiring decisions: A meta-analysis ofcorrespondence tests 1990-2015. J. Ethnic Migr. Stud. 42, 1115–1134 (2016).

36. K. Auspurg, A. Schneck, T. Hinz, Closed doors everywhere? A meta-analysis of fieldexperiments on ethnic discrimination in rental housing markets. J. Ethnic Migr. Stud.45, 95–114 (2018).

37. J. Hainmueller, D. Hangartner, Who gets a Swiss passport? A natural experiment inimmigrant discrimination. Am. Polit. Sci. Rev. 107, 159–187 (2013).

38. J. D. Tjaden, C. Schwemmer, M. Khadjavi, Ride with me: Ethnic discrimination, socialmarkets, and the sharing economy. Eur. Sociol. Rev. 34, 418–432 (2018).

39. R. Koopmans, S. Veit, Cooperation in ethnically diverse neighborhoods: A lost-letterexperiment. Polit. Psychol. 35, 379–400 (2014).

40. F. Winter, N. Zhang, Social norm enforcement in ethnically diverse communities. Proc.Natl. Acad. Sci. U.S.A. 115, 2722–2727 (2018).

41. R. J. Crisp, R. Meleady, Adapting to a multicultural future. Science 336, 853–855(2012).

42. E. L. Paluck, How to overcome prejudice. Science 352, 147 (2016).43. S. Bundesamt. Bevolkerung und Erwerbstatigkeit: Bevolkerung mit

Migrationshintergrund—Ergebnisse des Mikrozensus 2017 (Statistisches Bundesamt:Fachserie 1, Reihe 2.2, Wiesbaden, Germany, 2018).

44. M. L. Sands. Exposure to inequality affects support for redistribution. Proc. Natl. Acad.Sci. U.S.A. 114, 663–668 (2017).

45. L. Balafoutas, N. Nikiforakis, B. Rockenbach, Direct and indirect punishment amongstrangers in the field. Proc. Natl. Acad. Sci. U.S.A. 111, 15924–15927 (2014).

46. L. Balafoutas, N. Nikiforakis, B. Rockenbach, Altruistic punishment does not increasewith the severity of norm violations in the field. Nat. Commun. 7, 13327 (2016).

Choi et al. PNAS | August 13, 2019 | vol. 116 | no. 33 | 16279

Dow

nloa

ded

by g

uest

on

Oct

ober

9, 2

020